[Frontend] Use request id from header (#10968)

Signed-off-by: Joe Runde <Joseph.Runde@ibm.com>
2024-12-09 22:46:29 -07:00
parent 391d7b2763
commit 980ad394a8
8 changed files with 27 additions and 13 deletions
--- a/vllm/entrypoints/openai/serving_chat.py
+++ b/vllm/entrypoints/openai/serving_chat.py
@@ -176,7 +176,8 @@ class OpenAIServingChat(OpenAIServing):
            logger.exception("Error in preprocessing prompt inputs")
            return self.create_error_response(str(e))

-        request_id = f"chatcmpl-{request.request_id}"
+        request_id = "chatcmpl-" \
+                     f"{self._base_request_id(raw_request, request.request_id)}"

        request_metadata = RequestResponseMetadata(request_id=request_id)
        if raw_request: